Testing

Overview

Skynet MCP uses Bun's native test runner for all testing. The test suite covers all tool categories with both unit and integration tests.

Running Tests

Run all tests

bun test

Watch mode (auto-rerun on changes)

bun test:watch

Run specific test file

bun test tests/infrastructure.test.ts

Run with coverage (if configured)

bun test --coverage

Test Structure

tests/
├── infrastructure.test.ts  # Docker stack management
├── database.test.ts        # Memgraph & ChromaDB access
├── memory.test.ts          # Semantic memory operations
├── interactions.test.ts    # User interaction tracking
├── cognitive.test.ts       # Skynet workflow
└── integration.test.ts     # End-to-end scenarios

Writing Tests

Basic Test Structure

import { describe, expect, test } from "bun:test";
import { myFunction } from "../src/tools/myfile";

describe("MyFeature", () => {
  test("should do something", () => {
    const result = myFunction();
    expect(result).toBe(expectedValue);
  });
});

Mocking Databases

import { describe, expect, test, beforeAll, afterAll } from "bun:test";

describe("DatabaseTests", () => {
  beforeAll(async () => {
    // Setup test database
  });

  afterAll(async () => {
    // Cleanup test database
  });

  test("should query data", async () => {
    // Test implementation
  });
});

Integration Tests

Integration tests require running Docker containers:

import { describe, expect, test } from "bun:test";

describe.skipIf(!process.env.RUN_INTEGRATION)("Integration Tests", () => {
  test("full workflow", async () => {
    // Test requires Memgraph + ChromaDB running
  });
});

Run with:

RUN_INTEGRATION=1 bun test

Test Guidelines

Unit Tests

• Test individual functions in isolation
• Mock external dependencies (databases, Docker)
• Fast execution (< 100ms per test)
• No side effects

Integration Tests

• Test tool interactions end-to-end
• Require running databases
• Slower execution acceptable
• Use describe.skipIf() to skip when containers unavailable

Best Practices

• Descriptive names: Use clear test descriptions
• Arrange-Act-Assert: Structure tests clearly
• Independent tests: No shared state between tests
• Cleanup: Always clean up test data
• Edge cases: Test error conditions and boundaries

Current Test Coverage

Category	Tests	Status
Infrastructure	3	✅ Passing
Database	4	✅ Passing
Memory	2	✅ Passing
Interactions	6	✅ Passing
Cognitive	3	✅ Passing
Integration	2	⏭️ Skipped (requires containers)

Continuous Integration

Tests run automatically on:

• Every commit (via Git hooks, if configured)
• Pull requests
• Main branch pushes

See .github/workflows/ for CI configuration.

Debugging Tests

Verbose output

bun test --verbose

Inspect failures

bun test --bail  # Stop on first failure

Log debugging

test("debug example", () => {
  console.log("Debug info:", someValue);
  expect(someValue).toBe(expected);
});

Next Steps

• Building for Production
• API Reference
• Contributing Guide